INDEX

Explanations

the word "here" used to introduce an explanation or point

oai_token-act-pair · gpt-3.5-turbo

the phrase "here" indicating points of emphasis or important information

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.hook_resid_pre

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

torch.float32

Hook Point

blocks.4.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

omore

-0.70

itialized

-0.66

visors

-0.64

 hygiene

-0.60

 Heights

-0.60

iven

-0.58

ONSORED

-0.57

 tongues

-0.57

utter

-0.56

ggle

-0.55

POSITIVE LOGITS

abouts

1.44

tics

1.40

tical

1.38

tic

1.31

here

0.96

 guiActiveUn

0.90

fires

0.77

with

0.74

ridges

0.69

lander

0.68

Activations Density 0.044%

the word "here" used to introduce an explanation or point

the phrase "here" indicating points of emphasis or important information

No Comments

No Known Activations